Home

Awesome

The challenge

Challenge 1 (easy): modify the code so that it works for any resolution (width, height), e.g. 640 x 480.

Challenge 2 (medium): implement Bridson's Poisson disk sampling algorithm with other Python packages, use the same config with this repo (400x400 grid and 100K desired points) and whatever acceleration tricks. See if you can beat the speed of Taichi. (the compile time will not be counted)

Challenge 3 (hard): improve the code so that the result is a maximal Poisson sampling, that is, there won't be any room left to insert new points.

Please submit your work in this issue.

Install taichi

pip3 install -r requirements.txt

An interative animation

Mouse and keyboard control:

  1. Click mouse to choose an initial point.
  2. Press p to save screenshots.

Example:

<p align="center"> <img src="./demo.jpg" width="400" ></img> </p>

Benchmark with NumPy and Numba

See this repo:

https://github.com/taichi-dev/taichi_benchmark/tree/main/poisson

You are encouraged to implement a faster one to beat ours!